Dependencies

read in data

Data compiled in the prepDataForModels.R script

Here are the climate variables we could potentially use in the models

variable unit
swe_meanAnnAvg_30yr mm
tmean_meanAnnAvg_30yr degrees C
prcp_meanAnnTotal_30yr mm
precip_Seasonality_meanAnnAvg_30yr coef. of variation
PrecipTempCorr_meanAnnAvg_30yr correlation coef.
isothermality_meanAnnAvg_30yr coef. of variation
annWaterDeficit_meanAnnAvg_30yr mm of water/degrees celsius
annWetDegDays_meanAnnAvg_30yr degree days
annVPD_mean_meanAnnAvg_30yr KPa

Correlation of potential predictors

This is using a subset of the data (50,000 rows), just for runtime purposes

Below is the correlation between only climate predictors that are averaged across 30 years * Dropped tmin, tmax, t_warmest month, and t_coldest month and replaced w/ MAT * Drop prcp wettest month - use MAP only * dropped precip of driest month–was highly correlated w/ precip and precip seasonality * drop prcpSeasonality – highly correlated w/ water deficit and wet degree days * Replace VPDmean with VPDmax * Probably drop FreezeMon ** also dropped vp (which we didn’t talk about, but is highly correlated w/ VPD and pretty highly correlated w/ precip)

Visualization of some predictors

spatial distribution of VPD_max

spatial distribution of Wet degree days

spatial distribution of water deficit

visualize the residuals from the ‘best mod’ for total tree cover

Correlations at Ecoregion scales

Here are the ecoregions